HHAI 2026 Workshop · July 6–7

Question Zero

A workshop on responsible pre-deployment decisions for AI systems

"Under what conditions is it responsible to build and deploy an AI system for this problem?"

Submission deadline

May 11, 2026

Notification

May 15, 2026

Workshop

July 6–7, 2026

Submit an abstract →

Overview

Why? Asking before building

Many failures of deployed AI systems do not stem from technical limitations, but from premature or unjustified decisions to apply AI in the first place. This workshop introduces Question Zero (Q0) as a foundational inquiry into AI design.

We position Q0 as a pre-deployment safety constraint, shifting attention from downstream fixes such as interpretability, auditing, alignment patches, to upstream design decisions: problem formulation, system scope, human roles, and institutional context.

The workshop focuses on hybrid human–AI systems, emphasising how early choices shape technical feasibility, ethical outcomes, and societal impact. We ask not just "can we build this?" but "should we?"

Day 1 — Academics & Practitioners

AI researchers Practitioners Social scientists & ethicists

Day 2 — Policy & Regulation

Policymakers & regulators Legal scholars

Topics of Interest

What we're exploring

Question Zero (Q0)

Formalizing pre-deployment justification frameworks

Problem Framing & Scope

When is AI appropriate? Boundaries of beneficial application

Human–AI Task Allocation

Oversight models, responsibility distribution, and control

Institutional & Governance Context

Accountability structures and regulatory frameworks

Reversibility & Exit Strategies

Withdrawal conditions, rollback, and sunset planning

Lifecycle Implications

Constraints on auditing, explainability, and maintenance

Evaluation of AI Proposals

Criteria beyond technical novelty and performance metrics

Policy & Regulation

Translating Q0 reasoning into governance and law

Security Perspectives

Q0 from a security and threat-model lifecycle view

Case Studies

Real-world failures and successes in AI deployment decisions

Topics above are suggestions. We strongly encourage submissions to address Q0 directly for their specific case.

Workshop Format

Highly interactive, by design

Day 1 · July 6

Academic & Practitioner Focus

Full day · For researchers, practitioners, and framework contributors

Identify challenges and opportunities in the pre-deployment space of responsible AI. Contributions will feed into a structured synthesis and survey paper, including policy recommendations.

Lightning talks — 5 minutes per contribution
Group-based analysis of real-world cases
Collaborative development of Q0 frameworks
Structured discussions and synthesis sessions

Day 2 · July 7

Policy Roundtable

Half day · For policymakers, regulators, and policy-oriented researchers

Engage policymakers on how to target AI governance before technical questions are asked. Explore actionable, operational criteria for pre-deployment AI assessment and the tension between AI-first approaches and ethical and societal concerns.

Roundtable discussion with policymakers and regulators
Reactions and challenges from the policy perspective
Translation of research into operational policy criteria

Call for Papers

How to contribute

We invite 2-page extended abstracts offering a narrow, situated perspective on Question Zero. All submissions are reviewed for relevance, clarity, and contribution to workshop themes.

Case studies of AI deployment failures or successes
Conceptual or methodological frameworks
Policy-oriented analyses
Empirical or interdisciplinary work

Synthesis paper: All contributions will feed into a collaborative survey and structured synthesis of the pre-deployment AI space, including policy recommendations.

Proceedings: Authors may also opt to contribute to a CEUR-WS open-access proceedings volume.

Link to EasyChair to be published. Submit an abstract →

Important dates

May 11Submission deadline
May 15Acceptance notification
Jul 6–7Workshop at HHAI 2026

Organizing Committee

Who we are

Rachele Carli

Umeå University

Adam Dahlgren Lindström

Umeå University

Virginia Dignum

Umeå University

Bertilla Fabris

Umeå University

Maja Fjaestad

Karolinska Institutet

Janet Rafner

University of Southern Denmark

Jacob Sherson

Aarhus University

Tatjana Titareva

Umeå University

Jason Tucker

Institute for Futures Studies & Umeå University

Motivation

Why this workshop?

While much of AI research focuses on improving systems after they are built, this workshop asks a more fundamental question. By addressing Q0 before systems are deployed, we can prevent harm before it occurs rather than patching it afterward.

Policymakers and regulators need tools to operationalize "responsible AI" beyond high-level principles. This workshop offers early-stage decision criteria that bridge the gap between ethical commitments and actionable governance.

Prevent harmful or unnecessary AI deployments
Improve alignment with real-world human needs
Promote responsible, human-centred AI design
Bridge AI ethics principles and operational policy criteria

Preliminary material

Explore the Q0 assessment tool developed ahead of this workshop — a practical resource for reasoning through pre-deployment decisions.

aipolicylab.se/assessment-tool →